Face, Detected: Is Artificial Intelligence (AI) A Threat to Fashion Models?

On the call sheets: AI-powered models.
Ai models fashion

If you lay out the fashion calendar, there are four major fashion weeks: New York, London, Milan, and Paris. There’s also Haute Couture, the inter-seasonals like Cruise and Pre-fall, menswear, and RTW shows held in cities that are equally, if not more, fashion-forward. Think Copenhagen, Shanghai, Seoul, Tokyo, and more.

Launchmetrics, an AI-powered brand performance tool, reported that physical fashion weeks were capable of increasing brand visibility, consumer engagement, and market positioning. These shows serve more than just being a showcase—they’re also a centre of creativity and culture, which is particularly important for emerging labels or brands undergoing rebranding. Despite that, there’s a spike in designers opting out from the traditional fashion week calendar and pivoting from the formatted seasons. Reasons range from logistics to environmental concerns. After all, being sustainable is very much encouraged among consumers today.

Which brings to mind the question: Do we need yet another fashion week? An AI Fashion Week, to be precise. During the lockdown, we had our fair share of watching fashion shows virtually. AI Fashion Week, however, is not the same–it’s a show generated wholly through AI. During its debut in April 2023, the fashion world was shaken by the jaw-dropping work of AI fashion designers. Showcased on the virtual runway were intricate details and otherworldly silhouettes that we could only dream of but not achieve with existing materials and skills, which were, to say the least, a fashion design milestone. While sceptical, AI-generated content is a celebration of creativity and creates more job opportunities. But we question, to what extent should the fashion industry fully embrace AI? Will this advancement eventually replace the team behind brands? Or are they really, as OpenAI founder Sam Altman says, “wildly overhyped”?


If you look up, “Will AI replace human jobs?” on the Internet, experts will tell you the answer is no. Data, however, shows otherwise. “The Future of Jobs Report 2020” by the World Economic Forum stated that AI will take over an estimated 85 million jobs worldwide. However, the good news is that it is also creating 97 million new jobs simultaneously. It’s safe to say that while AI will not take over humans completely, it will replace humans without AI knowledge—and that is a huge portion of society. In fashion, for example, the automation system is gradually phasing out manufacturing roles. As a tool to analyse consumer preferences and trend forecasts, however, it’s making great strides. Similarly, generative AI can be useful for fashion labels in developing new ideas or refining existing designs; but it may remove manual jobs like modelling.

The European Union recently established the world-leading AI Act, setting a standard for countries like Malaysia that have yet to enact specific legislation. A risk-based approach was adopted to regulate AI: the higher the risk, the stricter the rules. Up on the “unacceptable risk” is facial recognition–that is, our identity.

@shereenwu Michael Costello has yet to take responsibility for his actions. I want to explain what happened, and I hope other models in the future feel comfortable to speak up. He has since offered to post my photo side by side with the AI one, but has not voluntarily post it. This offer did not contain an apology, and only happened after a model who’s close with him brought light to the situation. Some points I couldn’t fit: -Replaced the face of a model from the same collection -Lightened the skin of a black model in a photo and proceeded to push blame on the makeup artist (mua did not lighten the skin) -Screamed at models backstage (he screamed at the girl who stumbled on the runway to near tears, only to make an Instagram post praising her after.) But considering how long he’s been acting like this, I doubt any apology from him would be sincere; they would be performative at best. (Sorry for the weird cuts and sped up video I was trying to fit as much as I can in) #michaelcostello #greenscreenvideo #drama #michaelcostellocontroversy ♬ original sound – shereenwu

Identification comes in various forms, but before we delve into the creative identity, let’s start with the obvious ones like our faces, which can now be eerily altered or replaced. Several months back, Michael Castello, a Los Angeles-based designer favoured by A-listers like Beyoncé, Lady Gaga, and Celine Dion, was found to be deliberately whitewashing Asian-American model Sheeren Wu. She later exposed the designer on her TikTok page. In her video, Wu lamented that she understood how models, especially unpaid ones and not tied to an agency like herself, are replaceable. “But editing my face and removing my race is completely disrespectful,” the model said.

Tweaking models’ facial features and body shapes is not a new thing. Photoshopping models in editorials and commercials has been prevalent for decades. Wu’s incident, however, shed light on the potential that AI has in modifying and removing our digital footprints. In her case, it was crucial to her career development. In an interview with The Guardian, Susan Scafidi, academic director of Fordham’s Fashion Law Institute urged that the AI program is “turning back the clock” on the progress towards diversity, particularly in the fashion industry.

As it’s a machine-learning application, AI is fed with existing mainstream beauty preferences that are later analysed and permeated into the outputs that could potentially erase the identity of a real-life model. The legal loophole, Scafidi highlighter, is that the copyright law protects creators like photographers to a certain extent, should their work be altered without permission–but it does not apply to the models in the photos.

There’s a long road ahead in achieving diversity and inclusivity for both body shapes and racial justice. The ascent of generative AI, for better or for worse, has become a slam-dunk solution for many brands. Levi’s, for instance, partnered with Lalaland.ai to create AI-generated clothing models, or should we say, avatars, to tackle diversity and inclusivity issues. Artificial diversity, they claimed, is “allowing the customers to see our products on more models that look like themselves, creating a more personal and inclusive shopping experience”, read the announcement, except this personalisation can be done by hiring actual models with different body shapes. As a consumer, we feel human models are more convincing, making an actual impact on our buying decisions.

In a world where information is so accessible, consumers are equipped with a high level of awareness of body diversity and racial sensitivity. Thus, this decision to replace models–a profession that survives on minimum wage–with AI-generated models will backfire. The bigger picture, however, sees this step helping with operational costs. From a business point of view, it’s not hard to see why it makes more sense for brands to jump onto the AI bandwagon, replacing humans with AI-generated avatars.

“Will AI replace us?” is a frequently asked question, not just in fashion but also in many other fields. For an industry that revolves around creativity, it will take time to witness major disruptions in hiring. Take Collina Strada, for example, who showcased its AI-influenced Spring 2024 collection in New York, or Marc Jacobs’ Autumn/Winter 2023 show notes that were written by ChatGPT. It shows that AI is currently a mere wow factor on runways, but fashion is a business at its core. The widespread adoption of AI will eventually happen and eliminate some jobs, more likely entry-level roles such as pattern cutting, copywriting, modelling, or even designing. After all, the rich only get richer, and the poor get poorer.